C-LARA

This week, I have mostly been working on the ComputEL-7 paper, “reading histories” and better support for Mandarin:

ComputEL-7 paper

Pauline and I completed the final version of the ComputEL-7 paper and sent it in. The result was a great deal better than the rather sketchy submission version, we made good use of the week and a half they granted us.

‘Reading histories’

We now have a basic first version of ‘reading histories’ installed on the server. You can define your personal reading history for a given language, add texts to it, and see an automatically created multimedia result which consists of the concatenation of the texts including a joint concordance.

So far, the reading histories don’t include phonetic texts, which we obviously want. I am working on it and should be able to include that functionality soon.

Better support for Mandarin

When we discussed using C-LARA for Mandarin last year, one of the key items the Mandarin-speakers stressed was support for pinyin (Roman-alphabet glosses for Mandarin words). I have now added a first version using an integration of the popular pypinyin package; to my surprise, this apparently does better than GPT-4, which is available as an alternative in case we want to experiment.

Check out an example of a Mandarin C-LARA text with pinyin here. Belinda and I both thought GPT-4’s poem in praise of cat food was hilarious, even if we couldn’t actually understand the Mandarin and had to appreciate it through the glosses!

Next Zoom call

Thu Feb 8, 2023, 20:00 Adelaide (= 09.30 Iceland = 09.30 Ireland/Faroe Islands = 10.30 Europe = 11.30 Israel = 13.00 Iran = 17.30 China = 20:30 Melbourne/New Caledonia)